Last Tuesday, OpenAI announced it had reinstated Sam Altman, the CEO the board had abruptly fired four days prior, and formed a new board for the company. After a weekend of chaos, the message seemed to be that OpenAI’s governance was getting back on track. 

But board experts who spoke to Fortune are not entirely convinced. They say that so far, OpenAI’s post-failed coup board seems underwhelming for at least three key reasons.

1. It lacks diversity. The new board now includes Adam D’Angelo, the CEO of Quora, who is the only incumbent director, as well as Bret Taylor, the former co-CEO of Salesforce and former chair of Twitter, and Larry Summers, the economist and former head of the U.S. Treasury.

While Taylor and Summers are heavy hitters, what was an already all-white board is now also all-male. The two women who sat on the board last week—Helen Toner, director of Strategy and Foundational Research Grants at Georgetown’s Centre for Security and Emerging Technology, and Tasha McCauley, CEO of GeoSim Systems—are gone.  

The company has called this an “initial” new board, so it’s likely still working to recruit more people—and several names of women in tech have been floated by tech watchers—but the company sent a damning message by not prioritizing diversity first. OpenAI did not return Fortune’s request for comment.  

2. It’s lacking expertise in some critical areas. The OpenAI board needs directors who are more steeped in AI and the debate around where it’s going, says Greg Lau, who leads the board advisory practice at RSR Partners. “I think Larry Summers and Bret Taylor will bring a sense of order,” he says, “but they need some people that really understand what’s playing out there.”

The company also requires legal and cybersecurity know-how, and directors with more governance experience, especially considering the complicated structure of this company, says Patricia Lenkov, CEO and founder of Agility Executive Search. OpenAI, she adds, is a global company operating in several different markets that all come with their own set of challenges. That’s why the board “needs perspectives from other parts of the world.”

3. It’s too small. Robust governance requires multiple voices who can bring in counterpoints and minimize the risk of groupthink, says Lenkov. Even if this board were diverse, it would still be too thin. 

Again, to be fair, the board will likely grow and it may eventually include a seat for major investor Microsoft, The Information reported.

What comes next?

The next moves OpenAI’s directors need to make are clear, according to Lenkov and Lau: Diversify and expand the board, act like a cohesive team, map out a better communications strategy between the board and the CEO, as well as the company and outside stakeholders, and get to work on a proper CEO succession plan.

“T​he board must present a unified front and communicate their unequivocal support for Sam Altman,” says Lenkov. “Instability at the top of the organization is debilitating, so employees need to feel comfortable that they can get back to their work and not wonder about the leadership and direction of the business,” she says.

Succession plans also need to be put in place to avoid key-person risk or the possibility that a company would lose momentum, market share, and high-performing employees if one person were to leave the firm—something that became a clear liability when Altman was initially fired. “There must be not only a Plan B but a Plan C as well,” says Lenkov. OpenAI’s mission is “simply too important not to focus on this aspect of risk management,” she adds.

And finally, the board needs to be clear going forward with Altman about “what he’s doing, what his vision is, what his policies are,” says Lau. But Altman also likely needs to adjust his approach, he adds, and begin treating the board like a “true partner.” 

“I don’t know what it’s like dealing with Sam,” says Lau, “but his ego has probably gone up a few notches.”

Lila MacLellan
lila.maclellan@fortune.com
@lilamaclellan

Noted

“[T]he people who are building AI technology are not motivated by the things most business people are. This is less an industry than a religious movement. And I say that with no condescension. They don’t approach advanced artificial intelligence as someone who went to business school would, as a cost-benefit calculation. They have feelings that are deeply ideological and personal. Some of them think they are building a new God.”

—Kevin Roose, a New York Times tech reporter, explained what he has learned reporting on AI and OpenAI’s crisis, in a Times interview.

In Brief

—The Strategic Organizing Center—a coalition of labor unions including the Service Employees International Union—has nominated three director candidates for Starbucks’ board, the Wall Street Journal reports. The group, owner of a small stake in Starbucks, asserts that the board has failed to respond appropriately to unionization efforts at the coffee shops, thus exposing the retailer to “legal, financial and regulatory risks while damaging the value of its brand—and potentially its share price,” the Journal explains. 

—AI may be your new strategy partner, according to two American researchers in France who tested AI’s prowess for designing and working through business plans using INSEAD’s Blue Ocean framework, a deep strategy toolkit. In one instance, their AI software figured out that a probable headwind facing plans for a hypothetical new bagel shop in Paris is that tourists “may feel disappointed by the lack of availability, variety, and authenticity.”  

—Now the official sole head of Fox News following his father’s retirement, Lachlan Murdoch has inherited a major board lawsuit and other troubles. However, Bob Thompson at the Bleier Center for Television and Popular Culture at Syracuse University, told The Guardian: “What Lachlan Murdoch has to worry about most is still his dad.”

—Defined benefit pensions—as opposed to employee-funded and (usually) employer-matched defined contribution pensions—may be staging a private-sector comeback, the New York Times writes. Striking UAW workers named old-school pension plans among their demands this year, and even IBM has introduced a defined-benefit-style perk for its employees in what one expert calls a “back to the future” moment.